In Part I of this post, we explored some of the adaptations utilized in the tech space to help ease end users’ transition to new ways of doing things — Word documents that look like paper, email icons that look like envelopes, and the like. I noted that, while these adaptations are mostly harmless, their use in information governance has impeded adoption of more effective measures. Namely, when all information was on paper, storing that information in a containerized framework made sense. Searching for a needle in a haystack is much easier when you know which haystack to turn to. We closed with the following observation: strategies that make it easier for people to find information are not applicable when people are using computers to find the information.
The fundamental difference underlying those two processes is, of course, that computers process information in a different manner than people. A string of 0s and 1s is what actually underlies the intelligible text that people read on their screen. As such, containerizing digital information does not aid the computer in its task of finding the information requested by the end user; even worse, it directly impedes it. Imagine if, instead of having multiple disparate, containerized applications, an organization had one, giant container, upon which any desired function could be performed on any subset of the information contained therein, or on all of the information as a whole.
This may initially sound counterintuitive, because, to put it simply, we’re used to thinking in paper. Putting all of a company’s paper documents in a single container would make it harder — potentially impossible — to find the needle in the haystack. But computers are different: they can handle it. In fact, creating a data “lake” such as the one described above will make it easier for computers to do what they need to do. Such a framework allows for all of the data to be processed in a way that makes all of it mutually congruous and intelligible, and subsequently able to be searched and intelligently mined by one centralized, consistent search engine, as well as retained within the organization according to policies established at a single point of policy control.
Such a framework ensures that each and every search for information conducted within an organization is conclusive; that is, it can be known with complete certainty that each and every piece of information in the enterprise that is actually responsive to a given search query will appear in the results of that query. This has significant implications for the entire organization. For example, in the context of civil litigation, sanctions have been levied, and cases have even been won or lost, on the sole basis of organizations not being able to produce comprehensive data during discovery.
Decontainerized, unified information governance can not only aid in responding to today’s pressing issues; it can also lay the groundwork for transforming information stores from cost centers to profit centers. Unstructured data analytics can allow the enterprise to mine its information for insights into social connections, expertise, and performance, as well as to conduct real-time investigatory analysis. Instead of relying on potentially-biased subsets, or samples, of enterprise data, a unified framework enables analysis of the entire population, leading to conclusions whose statistical robustness are unimpeachable.
Containerized information architecture within the enterprise walls is anathema to effective information governance. While containerization is occasionally useful for front-end functions (think secure, “sandboxed” enterprise apps on a BYOD smartphone that can be wiped remotely), the information still needs to be managed under singular architecture on the back end. But the point remains: some analogies in the tech world are, at worst, harmless throwbacks to simpler times, while data siloes are actively and consistently interfering with businesses’ efforts to maximize the value they derive from their data.